CommE 5045, Machine Learning and Having It Deep and Structured, 2018 Spring, National Taiwan University (NTU)
- HW1: Several experiments on deep learning, including topics: deep vs shallow, optimization, and generalization.
- HW2: Sequence-to-sequence model, including two tasks: Video Captioning (hw2-1) and Chatbot (hw2-2).
- HW3: Generative Adversarial Network (GAN), including GAN and several tips implementation(hw3-1) and conditional GAN (hw3-2).
- HW4: Deep reinforcement learning, including Policy Gradient (hw4-1) and Deep Q Network (hw4-2).